Substituent Effects by Deuterium and Alkyl Groups ... - ACS Publications

Mitsuo Kira, Mieko Watanabe, and Hideki Sakurai*. Contribution from the Department of Chemistry, Faculty of Science, Tohoku University,. Sendai 980, J...
1 downloads 0 Views 740KB Size
5202

J . Am. Chem. SOC. 1980, 102, 5202-5207

Substituent Effects by Deuterium and Alkyl Groups and 13C Hyperfine Coupling Constants of Cyclopentadienyl Radicals as Studied by Electron Spin Resonance' Mitsuo Kira, Mieko Watanabe, and Hideki Sakurai* Contribution from the Department of Chemistry, Faculty of Science, Tohoku University, Sendai 980, Japan. Received December 4, 1979

Abstract: ESR spectra of the parent and six substituted cyclopentadienyl radicals (RCsH4.; R = H, D, Me, Et, l-Pr, t-Bu, Me3SiCH2),generated from the corresponding RCSH4SnMe3by an SH2attack on Sn with the photochemically generated tert-butoxy radical, have been recorded over the temperature range -95 to +24 OC. Judging from the proton coupling constants, the electron-releasing perturbation of alkyl groups is in the order (CH3)3SiCH2> CH3 > C2H5 > (CH3)2CH> (CH3)3C. Preferred conformations of these alkyl groups are discussed from the temperature effects on the spectra. Notably, the methyl group rotates freely, while the (CH,),SiCH, group has a fixed conformation in which the (CH3)3Sigroup eclipses the p orbital on the CI atom of the cyclopentadienyl. The resonance-integralperturbation effect is more important than the Coulombintegral perturbation in the deuterium substitution. Finally the coupling constants of 13Cfor C a n (n = 5-8) radical species are discussed.

Although a number of monosubstituted cyclic A hydrocarbon radicals have been investigated by ESR spectroscopy,2 there has been no such study on the substituted cyclopentadienyl radical until our report on several silyl-substituted cyclopentadienyl radicals appearedU3 This work revealed a unique role of the cyclopentadienyl radical as a tool for elucidating substituent effects because the unpaired electron of the cyclopentadienyl radical resides in a bonding A orbital, while in other familiar A radicals such as benzene anion and cycloheptatrienyl radicals, the unpaired electron occupies an antibonding orbital. We now report results of an ESR study involving deuterio- and alkyl-substituent effects as well as the I3C hyperfine coupling constant of the cyclopentadienyl radical. The required substituted cyclopentadienyl radicals were generated by a novel method in which a mixture of a substituted (cyclopentadieny1)trimethylstannane and di-tert-butyl peroxide is photolyzed at low temperature. We discuss particularly the Baker-Nathan effects and the preferred conformations of alkyl groups, enhanced u(C-Si)i conjugation of the (trimethylsily1)methyl group, the significant role of the out-of-plane vibration on deuterium isotope effects, and a rationalization of the I3C coupling constants of a series of C,H, (n = 5-8). Results and Discussion Generation and ESR Spectra of Substituted Cyclopentadienyl Radicals. The parent cyclopentadienyl radical was generated for the ESR study by several routes such as y irradiation on both c r y ~ t a l l i n eand ~ ~ ~liquid cyclopentadiene: pyrolysis of a molecular beam of ferrocene and azobenzene,' and hydrogen abstraction of cyclopentadiene with the tert-butoxy radical in solution.* The last method appeared to afford a general method of generating the substituted cyclopentadienyl radicals such as alkyl-substituted ones in solution. However, generation of substituted cyclopentadienyl radicals by this method is not always easy except for (1) Chemistry of Organosilicon Compounds. 131. For a preliminary communication, see: Kira, M.; Watanabe, M.; Sakurai, H. Chem. Lett. 1979, 973. (2) (a) Bowers, K. W. "Radical Ions"; Kaiser, E. T.,Kevan, L., Eds.; Wiley: New York, 1968; p 211. (b) Gerson, F.;Hammons, J. H. "Nonbenzenoid Aromatics 11"; Snyder, J. P., Ed.; Academic Press: New York, 1971; p 81. (3) Kira, M.; Watanabe, M.; Sakurai, H. J . Am. Chem. SOC.1977, 99, 7780. (4) Ohnishi, S.;Nitta, I. J . Chem. Phys. 1963, 39, 2848. (5) Liebling, G.R.; McConnell, H. M. J. Chem Phys. 1965, 42, 3931. (6) Fessenden, R. W.; Ogawa, S. J . Am. Chem. SOC.1964, 86, 3591. (7) Zandstra, P. J. J . Chem. Phys. 1964, 40, 612. (8) (a) Krusic, P. J.; Kochi, J. K. J . Am. Chem. SOC.1968, 90, 7155. (b) Kawamura, T.;Kochi, J. K. J. Organomer. Chem. 1973, 47, 79. (9) Engelking, P. C.; Lineberger, W. C. J. Chem. Phys. 1977, 67, 1413.

silyl- and germyl-substituted ones.3*10Additional signals due to an allylic radical result from addition of the tert-butoxy radical to the parent and alkyl-substituted cy~lopentadienes.~~~ Moreover, steady-state concentrations of these cyclopentadienyl radicals are intrinsically low. Therefore, detailed investigations of ESR spectra of these radicals under these conditions are very difficult (eq 1).

D

R R

oddltlon addltlon

tt t.Bu0. t.Bu0.

4 1

OBU-t

(1)

H absiraciion H absiraciion

We have found that photolysis of a mixture of an alkyl-substituted (cyclopentadieny1)trimethylstannane and di-tert-butyl peroxide at low temperature allows highly selective production of the corresponding cyclopentadienyl radical in high steady-state concentration through an SH2attack of the tert-butoxy radical at Sn.' The parent cyclopentadienyl radical including C S H ~ D ~ can also be generated in sufficiently high concentration to allow detailed analysis of the spectra (vide infra). We have thus generated parent (2a) and six substituted cyclopentadienyl radicals (2b-g) by this novel method (eq 2). Photochemically excited (CH3)3COOC(CHd3

1

2(CH313CO*

(2)

2a,R=H b,R=D c, R = CH, d, e, R = C,H, (CH,),CH

f, R = (CH,),C

g, R = (CH,),SiCH, acetone can also be used for the production of these radicals but without any improvement of the signal intensity. Very recently, Davies and his co-workers have generated the cyclopentadienyl radical by direct photolysis of d-C,H,,Sn'" and related organometallic species.'la In our hands, photolysis of (10) (a) ESR spectra of (trimethylgermy1)cyclopentadienyland other related radicals will be published in a forthcoming paper. (b) CI$i group, one of the most electron-accepting silyl group, gave the ul* of -0.56 G at -105 OC. (11) (a) Davies, A. G.; Tse,M. W. J. Chem. Soc., Chem. Commun. 1978, 353. (b) Barker, P. J.; Davies, A. G.; Fisher, J. D. Ibid. 1979, 587. ( c ) Professor Davies has informed us that they have investigated independently the ESR spectra of deuterio- and several alkylcyclopentadienyl radicals generated by their method. (d) Barker, P. J.; Davies, A. G.; Tse, M.-W. J . Chem. SOC.,Perkin Trans. 2, to be submitted for publication.

0002-7863/80/ 1502-5202$01 .OO/O 0 1980 American Chemical Society

J . Am. Chem. SOC.,Vol. 102, No. 16, 1980 5203

Coupling Constants of Cyclopentadienyl Radicals la)

O C H 1

Figure 3. ESR spectra of methylcyclopentadienyl radical (2c) at -35 (a) experimental; (b) computer simulated. (aI

OC:

0 C H 2 C H 3

Figure 1. ESR spectrum of the cyclopentadienyl radical (h) at -1 13 O C . 0

0 -125

-c

5 G

1 bl

Figure 4. ESR spectra of ethylcyclopentadienyl radical (Zd) at -125 OC: (a) experimental; (b) computer simulated. -92'c

I

I

5

G

Figure 2. ESR spectrum of the monodeuteriocyclopentadienyl radical (2b) superimposed with 2a at -92 O C . Table I. ESR Parameters of Substituted Cyclopentadienyl Radicals radi- temp, cal "C a,*a

-92 -95 -50 24 2d -95 -50 24 2e -95 -50 24 2f -50 24 2g -95 -50 24 2b 2~

a a,

5.86 13.04 12.98 12.72 12.74 12.80 12.28 12.54 12.66 12.24 12.38 12.20 14.04 14.00 14.06

* = QH -

hfcc/G a2,5

a,

other

6.19 5.99 a D = 0.90 0.76 7.82 15.29 0.82 7.79 15.14 1.00 7.73 14.80 0.94 7.79 15.99 1.00 7.70 15.33 1.24 7.72 14.50 1.12 7.71 12.56 ap=0.41b 1.15 7.62 12.28 0.42 1.30 7.68 11.78 0.41 1.30 7.61 ap = 0.67 1.48 7.52 0.63 0.50 7.58 10.40 0.48 7.62 10.49 0.56 7.51 10.57 ai (QH= 30.20 G; see this text).

RC 1.17 1.17 1.17 1.26 1.20 1.18 1.00 0.97 0.96

Figure 5. ESR spectra of isopropylcyclopentadienylradical (Ze) at -33 OC: (a) experimental; (b) computer simulated.

5 G

5'C

0.74 0.75 0.75 Figure 6. ESR spectra of tert-butylcyclopentadienylradical (20 at 5 "C: (a) experimental; (b) computer simulated.

0

101

trimethyltin species in the presence of di-tert-butyl peroxide gave stronger ESR signals. The ESR spectrum of 2a, as shown in Figure 1, was dramatically improved over that obtained by the previous route (eq 1).8 Highly selective production of 2a with no additional paramagnetic species such as a tert-butoxy adduct radical allowed the first observation of the satellite spectrum due t o ~yclopentadienyl-'~C radical in natural abundance. Since 2b was produced from the mixture of (deuteriocyclopentadieny1)trimethylstannane and (cyclopentadieny1)trimethylstannane (ca. 1:1), the ESR spectrum was superimposed with that of 2a as shown in Figure 2. The latter served as an internal standard for precise determination of the hyperfine coupling constants (hfcc). The ESR spectra of five alkylcyclopentadienyl radicals (2c-g) are shown in Figures 3-7.

tHISIICH,JI

5

-112'C

lbl

G

,I

,p ; i r e p " 111 ,I,

~,P, ,1

'

,-nLWq!

I

i I'

h I,

pi +-Up; ,

Y

*

Y

Figure 7. ESR spectra of [(trimethylsilyl)methyl]cyclopentadienyl radical (2s) at -1 12 O C : (a) experimental; (b) computer simulated.

All the ESR spectra of these radicals were analyzed with the aid of computer simulations. Table I lists the hfcc's and other ESR parameters of these radicals. Ring proton hfcc's were as-

5204

J . Am. Chem. SOC.,Vol. 102, No. 16, 1980

Kira, Watanabe, and Sakurai density p i at position i of a substituted cyclopentadienyl is given by eq 6 and 7, where Ps and PA are the relative probabilities that pi = PSPS

a-R

t

+ PAPiA

the radical is in the S and A forms, respectively, at a particular temperature and p i s and p? are the spin densities at position i in the S and A forms, respectively. According to the McConnell relationship, ai = Q H p i , 1 2 the a equivalent relation to eq 6 for the hfcc is deduced as eq 8 where ai is the proton splitting in the radical ai = p s a s

+ PAaiA

(8)

at position i and ais and a? are the proton hfcc's at position i in the S and A forms, respectively. Since the total *-electron spin density must be unity, the spin density at position 1 would give rise to an a-proton splitting al* defined by eq 9. 5

ai* =

Figure 8. HMO energy diagram for cyclopentadienyl radical.

QH -

Xai

i=2

In addition, one can evaluate P A and Ps from the measured values of ai's on the basis of both HMO and McLachlan calculations. Values of AE can be calculated from the temperature dependence of P,/PS (see eq 10). PA/Ps = exp(-AE/kT)

Yi

X = E l e c t r o n Acceptor

X Z E l e c t r o n Donor

X=H

bitals in the monosubstituted cyclopentadienyl radical.

signed on the basis of the prediction of the H M O theory. Following the previous p~blication,~ we discuss the substituent effects by the Huckel molecular orbital model without taking into consideration the effects of extensive configuration interaction and vibronic coupling. The Jahn-Teller effects for cyclopentadienyl radical are concluded to be rather weak by the investigation of the photoelectron spectra of C5H5-.9 This simple treatment is useful enough to draw a qualitative picture of the character of the substituent, especially for a series of closely related cyclopentadienyl radicals. The HMO Model for Substituted Cyclopentadienyl Radicals. Since the perturbation treatment of the substituent effects on the D5,,cyclopentadienyl radical was already been described in detail in the previous publication within the framework of Huckel molecular orbital t h e ~ r yonly , ~ the essential points necessary for the subsequent discussion are repeated. In a monosubstituted cyclopentadienyl radical with the local symmetry of Cb in which the substituent occupies the position 1, the unpaired electron is unequally distributed between the two orbitals 9 s and \kA which are no longer degenerate as they are in the parent radical (eq 3 and 4). These orbitals are depicted \ks = O.632Xl + O.195(X2 + X5) - O.512(X3 + ~ 4 ) (3)

+

= 0 . 6 0 2 ( ~ 1- ~ 5 ) O.372(~3- ~

q )

(4)

schematically in Figure 8. An electron-donating substituent raises (S form) the \ks level and the electronic configuration 'kAzPsl is preferred, while an electron-accepting substituent lowers the q s level and the 3s2\kA1configuration (A form) is preferred. Two types of perturbations on the degenerate highest occupied MO's of CSH5-are schematically shown in Figure 9. When the energy difference ( A E ) between the two configurations is small, appreciable mixing of the two forms occurs (eq 5 ) . The observed spectrum is a result of the rapid exchange \kA2\ksl

(10)

On the basis of the observed Q H value of the cyclopentadienyl radical (QH= 30.2 G) and HMO or McLachlan12bcalculation, the following spectrum of al* values is expected by systematically changing the electronic properties of the substituent.

Figure 9. Splitting of degeneracy of the highest ocupied molecular or-

'PA

(9)

\ks2\kA'

(5)

between S and A forms. According to this assumption, the spin

X = a strong donor X = H -+ X = a strong acceptor pure S form pure A form a,*=12.1 G a , * = O G (HMO) a , * = 14.5 G + a , = 6.04 G-+ a , * =-2.7 G (McLachlan, h = 0.75)

We have actually observed al* values in the range 4 . 5 6 to +10.7 G for substituted silyl group^.^.'^

Alkyl-Substituent Effects on C5H5.:13Baker-Nathan Effects and Enhanced u--A Conjugation of (Trimethylsi1yl)methyl Croup. All the alkyl-substituted cyclopentadienyl radicals (2c-f) showed a]* values larger than 12 G as seen in Table I. Since the al* value of the pure S form can be estimated as 12.08 G from the H M O spin density cal~ulation,~ one might judge that all the alkyl substituents raise the \ks orbital of C5HS-to make the energy difference between the ground S form and the excited A form large enough to force the S form population Ps to unity. We have observed, however, small but appreciable temperature dependences of al* values for 2c-f even for the methyl derivative where the substituent has the local symmetry of C3". On the other hand, the al* of 2g whose substituent is the strongest electron donor was rather larger than those due to extensive g - r c~njugation'~ of 2c-f and temperature independent. These observations suggest that the limiting value of the a ] * for the pure S form may be a little larger than 12.1 G predicted by the HMO theory, being taken (12) (a) McConnell, H. M. J . Chem. Phys. 19!%,24,764. (b) McLachlan, A. D. Mol. Phys. 1960, 3, 233. X = 0.75 was used so as to give the best fit

between the observed five-membered ring proton hfcc values for the indenyl radical and the calc~lation.~ (13) Substituent effects of methyl group on CsHS. were examined by photodetachment investigations of CH3CSH4-:Richardson, J. H.; Stephenson, L. M.; Brauman, J. I. J . Chem. Phys. 1973, 59, 5068. (14) (a) Kawamura, T.; Kochi, J. K. J . Am. Chem. SOC.1972, 94, 649. (b) Symons, M. C. R. Ibid. 1972, 94, 8589. (c) Tetrahedron Lerr. 1975,193. (d) Griller, D.; Ingold, K. U. J . Am. Chem. SOC.1973, 95, 6459. (e) [bid. 1974,96,6715. (f) Sakurai, H.; Uchida, T.; Kira, M. J . Organomef. Chem. 1976,107, 15. (9) Kira, M.; Sakurai, H. J . Am. Chem. SOC.1977,99, 3892. (h) Bock, H.; Kaim, W. Tetrahedron Lett. 1977, 2343. (i) Bock, H.; Kaim, W.; Rohwer, H. E. J . Organomet. Chem. 1977, 135, C14. ti) Chem. Ber. 1978, 1 1 1 , 3573. (k) Bock, H.; Kaim, W. Ibid. 1978, 1 1 1 , 3552, 3585. (I) Bock, H; Kaim, W; Kira, M.; Osawa, H.; Sakurai, H.J . Organomet. Chem. 1979, 164, 295.

J. Am. Chem. Soc., Vol, 102, No. 16, 1980 5205

Coupling Constants of Cyclopentadienyl Radicals more properly as 14.5 G based on the McLachlan calculations (A = 0.75).12bThe observed small differences among al* values of 2c-f at a definite temperature can be taken as the reflection of the difference of the electronic effects of alkyls. The order of electron-donating ability of the alkyl groups estimated by this method is as follows: Me > Et > i-Pr > t-Bu, being apparently in accord with the order of the Baker-Nathan effect.lS Recently, Baker-Nathan effects have often been explained by some nonvertical effect such as steric hindrance to s~lvation,'~-'~ since ionization potentials,16b.mand charge-transfer spectral9 of alkylbenzenes give the reverse order of the Baker-Nathan effects. However, it should be noted that the substituent effects on some cyclic x radicals such as benzene anion2I and cycloheptatrienyl radicalsZ2have also revealed stronger apparent electron-donating ability of the methyl group than the tert-butyl group as observed in the cyclopentadienyl system. Baker-Nathan effects may be caused in a certain situation by some vertical electronic interaction. The difference of electron-donating ability between methyl and tert-butyl groups can possibly arise from the more effective negative h y p e r c o n j u g a t i ~ nof~ the ~ ~ ~latter ~ ~ ~ ~than ~ the former, since the UT* splitting is probably smaller for a C-C than a C-H bond.23c As indicated by the a l * value of 14.0 G and the temperature independence, the [(trimethylsilyl)methyl]cyclopentadienylradical is considered to be in the pure S form in the experimental temperature range due to the extensive electron donation by u(CSi)-x c o n j ~ g a t i o n .The ~ ~ significantly larger al* of 2g than that of 2c is incompatible with the idea of homo (p-d) conjugation.14,d~e Conformational Preference of Alkyl Substituents on Cyclopentadienyl Radicals. From the molecular orbital consideration for cyclopentadienyl radicals one can expect that the electrondonating alkyl substituents make the \ks orbital the singly occupied molecular orbital (SOMO). The alkyl proton hfcc values on C,, 2

5

H

H'

defined as a, here, can be induced mainly by a hyperconjugation mechanism from the rather high spin density on the C1 px orbital, while a, values of cycloheptatrienyl and benzene anion radicals are fairly small since one of the nodal planes of the SOMO passes through the substituted C1atom. Relatively large a, values allow one to investigate the conformational preference of alkyl groups in XCSH4radicals. The magnitude of a, should be dependent on both the spin density at C1 and the time-averaged dihedral angle between the a-CH bond and the C1 px orbital, as represented by the following equationz4 ( A N 0):

+

a, = p l ( ~ B ( C O ~e )~)

= p l ~ ( c o s 2e )

(1 1)

(15) Baker, J. W.; Nathan, W. S . J . Chem. SOC.1935,1844. (16)(a) Schubert, W.M.; Craven, J. M.; Minton, R. G.; Murphy, R. B. Tetrahedron Lett. 1959, 194. (b) Schubert, W.M.; Gurka, D. F. J . A m . Chem. SOC.1969,91, 1443. (17)Clement, R. A,; Naghizadeh, J. N. J . Am. Chem. SOC.1959,81,3154. (18) Himoe, A,; Stock, L. M. J . Am. Chem. SOC.1969 92, 1452. Berwin, H. J.; Traylor, T. G. J . Am. Chem. SOC.1970, (19)Hanstein, W.; 92, 829. (20)Turner, D. W.Adu. Phys. Org. Chem. 1966,4 , 31. (21)(a) de Bore, E.; Colpa, J. P. J . Phys. Chem. 1967,71, 21. (b) Jones, M. T.; Metz, S.; Kuechler, T. C. Mol. Phys. 1977,33, 717. (c) Jordan, K. D.; Michejda, J. A.; Burrow, P. D. J . Am. Chem. SOC.1976,98, 1295. (22)(a) Farr, E.;Rim, Y . S.; Bauld, N. L. J . A m . Chem. SOC.1971,93, 6888. (b) Vincow, G.;Morrell, M. L.; Hunter, F. R.; Dauben, H. J., Jr. J . Chem. Phys. 1968,48, 2816.

(23)(a) Hoffmann, R.; Radom, L.; Pople, J. A,; Schleyer, P. v. R.; Hehre, W. J.; Salem, L. J . A m . Chem. SOC.1972,94, 6221 and references cited therein. (b) F'urins, D.; Karplus, M. J . Am. Chem. SOC.1968,90, 6275. (c) However, as suggested by a referee, it is not clear that there is a simple physical model capable of predicting the relative a,* values, since the a,* differences among Zc-f are quite small. For the series of radicals, the observed differences are likely due to the subtle contribution of several factors, including vibronic terms. (24)Heller, C.;McConnell, H. J . Chem. Phys. 1960,32, 1535.

Combining the above equation with the McConnell relationship, we introduce R values as defined by eq 12. The R values are R = a,/al* ( B / Q ~ ) ( c oe ~) ~ (12) also listed in Table I. Since the values of QH and B should be independent of the alkyl substitutent, the observed R values depend only on ( cos2 e ) . In other words, we can take the R value as an indication of the time-averaged conformation of the alkyl group. The methyl group in 2c gives a value for R (=1.17) which is constant over the entire temperature range. Hence, we conclude that the methyl group is freely rotating ((cos20) = ' / J and B/Qg = 2.34, in reasonable agreement with values for other radicals. A slightly but appreciably larger R value for 2d (R = Et) than for 2c as well as the negative temperature coefficient dR/dT of the former radical suggests that 2d is more stable in conformation A than in B. A similar conformational preference is observed

H

A

B

C

for the n-propyl radical.25 On the contrary, some other ethylsubstituted cyclic x radicals such as ethylcyclooctatetraene anion?6 p-ethylnitrobenzene anion,27ap,p'-diethylbiphenyl anion,27band p-ethylphenyl alkoxy nitroxide radicals'4f are known to prefer another conformation B. Such a remarkable difference in the preferred conformations may be rationalized in terms of the smaller interior angle of the regular pentagon than those of the hexagon and the octagon. The steric hindrance between the ortho hydrogens and the methyl group in conformation A, as suggested by Carrington and Todd,26may be reduced. As a consequence, conformation A is preferred by 2d. Here one may call to mind the delicate balance between C-H and C-C hyperconjugations. The preferred conformation of 2e (R = i-Pr) is rather complicated to decide since the temperature dependence of the R value was not monotonous as seen in Table I, although the smaller R value than that of 2c can be taken as the indication of the preferred conformation C for 2e, in accord with those for other isopropyl-substituted aromatic anion radicals.21c,28 A small and constant R value for 2g is another indication of the extensive u(C-Si)-n conjugation. R (for 2g)/R (for 2c),which means 2(c0s2 0) for 2g, is 0.64 and can be compared with the values for (trimethysily1)ethyl (0.66)14' and some other related radicalsz9 Removal of Degeneracy by Deuterium Substitution. Verification of the Resonance-Integral Perturbation Model. Until now, there have been many investigationsof deuterium isotope effects on cylic x radicals, but the splitting of the two degenerate electronic configurations has been observed, only for the benzene-d anion radical.30 The substitution of benzene cycloh e ~ t a t r i e n y land , ~ ~cyclooctatetraene ~ anion radical32with deuterium showed no appreciable effect on the residual proton couplings. Two types of perturbing effects for deuterium substitution on a cyclic x radical, Le., Coulomb-integral and resonance-integral perturbation, have been suggested p r e v i o ~ s l y . ~The ~ , ~ former ~ (25)Fessenden, R. W.; Schuler, R. H. J . Chem. Phys. 1963,39, 2147. (26)Carrington, A.; Todd, P. F. Mol. Phys. 1964,8, 300. (27)(a) Mckinney, T.M.; Geske, D. H. J . Am. Chem. SOC.1967,89, 2806. (b) Nemoto, F.; Ishizu, K. J . Phys. Chem. 1975,79, 1730. (28)(a) Nemoto, F.; Shimoda, F.; Ishizu, K. E d / . Chem. SOC.Jpn. 1975, 48, 2627. (b) Chem. Lett. 1974,693. (29)The complete list of these values is compiled in ref 14k. (30)(a) Lawler, R. G.; Bolton, J. R.; Fraenkel, G. K.; Brown, T. H. J . Am. Chem. SOC.1964,86,520.(b) Karplus, M.; Lawler, R. G.; Fraenkel, G. K. J . Am. Chem. SOC.1965,87, 5260. (c) Lawler, R. G.; Fraenkel, G. K. J . Chem. Phys. 1968,49, 1 1 26. (31)(a) Carter, M. K.; Vincow, G. J . Chem. Phys. 1967,47, 292. (b) Volland, W.V.; Vincow, G. J. Chem. Phys. 1968,48, 5590. (32)Carrington, A.; Longuet-Higgins, H. C.; Moss, R. E.; Todd, P. F. Mol. Phys. 1965,9, 187.

5206 3. Am. Chem. Soc., Vol. 102, No. 16, 1980

Out-of+lone Vibration

X= H

Kira, Watanabe, and Sakurai

lnductfve E f i e c t

Figure 10. Splitting of degeneracy of the highest occupied molecular orbitals in the deuterium-substituted cyclopentadienyl radical and benzene radical anion.

takes account of the fact that deuterium releases electrons more strongly than hydrogen which will perturb the Coulomb integral (a).The perturbation energy depends on the atomic coefficient at the substitution position (c,) and can be described by the following equation: A.t(a) = ~ : 8 ~ ~ @ ~

(13) where 8 / P 0 is the Coulomb-integral change by deuterium substitution and should be positive. The resonance-integral (@)model considers the out-of-plane vibration of hydrogen or deuterium, which determines the average angle between the p r orbital on the CI atom and the p?r orbitals on the two adjacent carbon atoms. The fl should decrease with increasing angle. The perturbation energy can be expressed as follows: where 8r,,+l'@o is the resonance-integral change between the two adjacent carbon atoms produced by deuterium substitution and should be taken as negative owing to the reduced amplitude of the vibration. As shown schematically in Figure 10, the HMO perturbation theory predicts that for the benzene anion radical both perturbations destabilize the symmetric orbital \ks', while the antisymmetric orbital \kA' is unchanged; one can expect the configuration should be more important for the benzene-d anion radical than the \kdl configuration, although the latter contributes considerably, owing to thermal excitation and vibronic coupling. Experimental results30 are reasonably explained by both perturbation models. On the other hand, for the cylcopentadienyl radical two models give rise to different situations. According to eq 13 and 14, the symmetric orbital \ks should be stabilized and destabilized by the resonance-integral and the Coulomb-integral perturbation, respectively; the former makes the \ks'?A' configuration more important than the \kA2\ksi,while the situation is inversed for the latter. Although considerable contribution from the excited electronic configuration is expected because of thermal mixing and vibronic coupling, y = al*/aH, where al* is a proton splitting at the deuterium-substituted position of Zb, and should be the proper criterion of the predominant perturbation. Thus it should be expected that resonance-integral perturbation is more effective if y is less than unity, while Coulomb-integral perturbation should be the major factor if y is greater than unity. The spectrum of Zb shows that the twofold degeneracy of the ground-state electronic configuration of the cyclopentadienyl radical is lifted by deuterium substitution. Thus, two sets of triplets (intensity ratio 1:2:1) of 6.19 f 0.02 G and 5.99 f 0.02 G, which are further split into a triplet (1:l:l) of 0.90 f 0.01 G, are observed. The ratio, r, is 0.969 i 0.008 from 12 measurements, which indicates that the resonance-integral perturbation is more important than Coulomb-integral perturbation. While the apparent ratio of magnetic moments of hydrogen to deuterium, uiH/uiD, in the benzene anion has been shown to (33) The absence of a significant deuterium isotope effect in C8H7D-*is explained by an alternative consideration. See ref 32.

Figure 11. ESR spectrum of the cycloheptatrienyl radical at -105 " C . Table 11. Experimental and Calculated I5C hfcc Values for Cyclic n Radicals (G) calcd radical

exptl (temppC)

K-FJ7

Y-KKSa INDO4I

C,H,. C,HL. C,H7. C,H8-.

2.66 f 0.02 (-105) 2.8 i 0.1 (-100) 1.98 * 0.02 (-105) 1.28 ?r 0.05

1.56 1.30 1.11 0.98

2.04 2.23 1.76 1.43

4.1 4.0 3.5 3.0

ref this work 42 this work 43

differ significantly from g H / g D (=6.514),29cal*/alD in 2b was found to be 6.48 f 0.08, bordering on 6.514 within the experimental error. We cannot account for such a difference between C6HSD--and C5H4D. since the exact source of the anomalous deuterium hyperfine coupling has not been known up to date.34 Although vibronic coupling should be taken into account for the quantitative estimation of the deuterium isotope effect,35one may obtain the qualitative feature by the rough calculation without inclusion of the vibronic coupling. According to the method of Volland and Vincow?lb the energy splitting (AE) between the two nearly degenerate states of C5H4D.can be calculated as follows: (15) AE = E, - E A = ~ R T A u ~ ~ / + ( AQ ~u( p~3 '~- ~ 2 ' ) ) where AaZ3is the hfcc difference between a2 and a3observed for C5H4D-and pis is the spin density at position i calculated for the S form. The value of AE for C5H4D. is 22 cal/mol which can be compared with 57 cal/mol for C~HSD-.and 25 cal/mol for C7H6D-.31bOn the other hand, one can expect the ratio of A@) (for C,H,D-.)/At(fl) (for CSH4D.) to be 1.4 from eq 14 under the assumption of the same 8fl@ofor both radicals. Therefore, the relatively small AE for C5H4D.may be attributed to the partial cancellation by Coulomb-integral perturbation. 13C Hyperfine Coupling Constant of C ~ H S - ~ ~The C . carbon-1 3 hfcc, a(I3C), of C5HS.was measured from the satellite lines in the spectrum of Za (see Figure 1) as 2.66 f 0.02 G at -105 "C. To compare the u(l3C) values of a series of cyclic T radicals at a low temperature, we have investigated the ESR spectrum of C7H7.3698a by generating the radical with the photolysis of the mixture of cycloheptatriene and di-tert-butyl peroxide at -105 "C. The ESR spectrum is shown in Figure 11. The estimated a(I3C) for C7H7-is 1.98 f 0.02 G , which may be compared with the value of 1.86 G extrapolated from the temperature dependence of the values determined at around 165 Table I1 lists the a(13C) values of pertinent cyclic r radicals together with the theoretical ones calculated by three different methods. The well-known Karplus-Fraenkel equation3' for a(l3C) is generally expressed as follows: (34) Jones, M. T.; Cairncross, A.; Wiley, D. W. J . Chem. Phys. 1965, 43, 3403. (35) Purim, D.; Karplus, M. J. Chem. Phys. 1975,62, 2 have calculated the deuterium isotope effects on the benzene anion radical, by taking account of the vibronic coupling to suggest the importance of the out-of-plane vibration for the vibronic energy. (36) Vincow, G.; Morrel, M. L.; Volland, W. V.; Dauben, H. J., Jr.; Hunter, F. R. J . A m . Chem. SOC.1965,87, 3527. (37) Karplus, M.; Fraenkel, G. K. J . Chem. Phys. 1961, 35, 1312.

Coupling Constants of Cyclopentadienyl R a d i c a l s

J . Am. Chem. Soc., Vol. 102, No. 16, 1980 5201

where SC = -12.7 G, Qcc, = 14.4 G, QCH = 19.5 G, and Qcc = 13.9 G are taken according t o t h e original proposal by t h e authors. For simple cyclic T radicals, this equation can be put in t h e following form:

Theory predicts that a(13C) should increase with decreasing ring size a s shown in t h e third column of Table 11. T h e Yonezawa-Kawamura-Kat0 treatment,3* an alternative method for estimating a(13C) from the ?r A 0 spin density matrix, is described by t h e following equation: 3

a("C) =

QCPCC

3

+ EQxpii + iC= l R C X ~ C ,

(18)

1=1

where Qc = 46.0 G, Qc, = -17.3 G, and Rcci = -1.95 G . For a simple cyclic ?r radical

T h e theory emphasizes t h e importance of off-diagonal elements of t h e ?r A 0 spin density matrix (pc), predicting smaller I3C hfcc of C5H5.t h a n t h a t of C6H6-' d u e to t h e positive pci value for SOMO of t h e former, a s shown in the fourth column of Table 11. This treatment gives rather good agreement with experiments, especially for t h e d r o p of t h e hfcc from C&-' to C5H5-. Differences of the bond angle39as well as the difference of excess charge@ between C5H5-and C6H6-' are not included in the above discussion. However, the influence of the bond angle may be small because a(13C), calculated by IND0,41which m a y be regarded a s a Karplus-Fraenkel type calculation including t h e bond angle effects, give t h e same order a s t h e Karplus-Fraenkel theory in the series of cyclic ?r radicals, as seen in Table 11. W h e n Bolton's theory@ of excess charge effects for the proton coupling constants is applied to t h e 13Ccoupling constant, a ( I 3 C ) for t h e benzene anion radical is modified to even a smaller value than t h a t evaluated by eq 17; thus t h e excess charge effect cannot account for t h e difference of a ( I 3 C ) between C5H5.a n d CsH6-..

I3c

Experimental Section Materials. Cyclopentadiene and methylcyclopentadiene were obtained by thermal depolymerization of the corresponding dimers. Cyclo(38) Yonezawa, T.;Kawamura, T.; Kato, H. J . Chem. Phys. 1969,50, 3482. (39) (a) Bernal, I.; Rieger, D. H.; Fraenkel, G. K. J . Chem. Phys. 1962, 37, 1489. (b) Higuchi, J. Ibid. 1963,39, 3455. (40) Bolton, J. R. J . Chem. Phys. 1965,43, 309. (41) Pople, J. A.; Beveridge, D. L. "Approximate Molecular Orbital Theory"; McGraw-Hill: New York, 1970; p 136. (42) Bolton, J. R. Mol. Phys. 1963,6,219. (43) Strauss, H. L.; Fraenkel, G. K. J . Chem. Phys. 1961, 35, 1738.

pentadiene-d was prepared with a D content of over 95% (analyzed by mass spectroscopy) by the treatment of CSHSMgBrwith D 2 0 in di-nbutyl ether according to the procedure of Mironov et al." Ethyl- and isopropylcyclopentadienewere obtained from CSH6and the appropriate alkyl bromide by using phase-transfer reagents.4s tert-Butylcyclopentadiene was prepared by the procedure described in the literature.46 [ (Trimethylsilyl)methyl]~yclopentadiene~~ was prepared from the reaction of sodium cyclopentadienide with (iodomethy1)trimethylsilane in T H F in 41.9% yield; bp 39-40 "C (3-4 mmHg). (terr-Butylcyclopentadienyl)trimethylstannane(If) was prepared by the reaction of t-BuCsH4Li with Me'SnC1 according to the method of Davison and Rakita" in 13% yield: bp 56-60 OC (2 mmHg); NMR (60 54.0 Hz), MHz, CC14) 6 0.082 (S, 9 H, JII?s,,